IDENTIFICATION OF LINEAR DYNAMIC SYSTEMS AS A STATISTICAL PROBLEM INVOLVING NUISANCE PARAMETERS

A. F. Kushnir

Abstract

We discuss the estimation of parameters of multivariate systems. The estimation is based on observations of input and output obscured by noise. Such problems arise in geophysical studies of seismic response of media and structures when experimental data are processed. We find lower bounds for estimator variance and design optimal estimation procedures. We also examine the stability of the estimator quality under variations of statistical noise properties.

Back to
Computational Seismology, Vol. 2.